Type: Package
Title: Core Collection
Version: 0.9.5
Description: Create a custom sized Core Collection based on a distance matrix and applying the A-NE (accession nearest entry), E-NE (entry nearest entry) or E-E (entry entry) method as introduced in Jansen and van Hintum (2007) <doi:10.1007/s00122-006-0433-9> and further elaborated on in Odong, T.L. (2012) https://edepot.wur.nl/212422. Optionally a list of preselected accessions to be included into the core can be set. For each accession in the computed core, if available nearby accessions are retrievable that can be used as an alternative.
License: GPL-2 | GPL-3 [expanded from: GPL (≥ 2)]
Encoding: UTF-8
Depends: R (≥ 3.5)
LinkingTo: Rcpp
Imports: Rcpp (≥ 1.0.0), R6 (≥ 2.4.0), methods
Collate: coreCollection.package.R coreCollection.R coreSelection.R RcppExports.R
Suggests: testthat (≥ 3.0.0), vcfR, adegenet, ggfortify
RoxygenNote: 7.2.2
NeedsCompilation: yes
Packaged: 2022-12-20 08:37:25 UTC; matthijs
Author: Matthijs Brouwer ORCID iD [aut, cre], Reinhoud Blok, de [ctb]
Maintainer: Matthijs Brouwer <matthijs.brouwer@wur.nl>
Config/testthat/edition: 3
Repository: CRAN
Date/Publication: 2022-12-20 13:00:02 UTC

The coreCollection package

Description

This package can be used to create a CoreCollection object.

Author(s)

Matthijs Brouwer <matthijs.brouwer@wur.nl>

References

Odong, T.L. (2012) Quantative methods for sampling of germplasm collections - Getting the best out of molecular markers when creating core collections. PhD diss., Wageningen University and Research, Wageningen, The Netherlands. http://edepot.wur.nl/212422

Jansen, J & Hintum, Theo. (2007) Genetic distance sampling: A novel sampling method for obtaining core collections using genetic distances with an application to cultivated lettuce. TAG. Theoretical and applied genetics. Theoretische und angewandte Genetik. 114. 421-8. 10.1007/s00122-006-0433-9

See Also

- vcfR provides a suite of tools for input and output of variant call format (VCF) files, manipulation of their content and visualization.
- adegenet provides the genlight class for genome-wide SNP data, and includes a method to create a distance matrix.

Other core collection: CoreCollection()

The CoreCollection Class

Description

The CoreCollection Class

The CoreCollection Class

Format

A R6Class generator object

Methods

Public methods

Method recompute()

Usage
.CoreCollectionClass$recompute()

Method alternativeCore()

Usage
.CoreCollectionClass$alternativeCore(n)

Method new()

Usage
.CoreCollectionClass$new(
  distanceMatrix,
  n,
  preselected,
  coreSelectMethod,
  adjustedGroupMethod,
  algorithm,
  seed
)

Method print()

Usage
.CoreCollectionClass$print(...)

Method summary()

Usage
.CoreCollectionClass$summary(...)

Method measure()

Usage
.CoreCollectionClass$measure(coreSelectMethod)

Method measures()

Usage
.CoreCollectionClass$measures(...)

Method clone()

The objects of this class are cloneable with this method.

Usage
.CoreCollectionClass$clone(deep = FALSE)
Arguments
deep

Whether to make a deep clone.

The CoreCollection class

Description

R6 class for creating a core collection based on the provided distanceMatrix, required size of the core n and optionally a set of preselected accessions to be included into the core.

Usage

CoreCollection(
  distanceMatrix,
  n,
  preselected = c(),
  coreSelectMethod = "A-NE",
  adjustedGroupMethod = "split",
  algorithm = "randomDescent",
  seed = NULL
)

Arguments

distanceMatrix

A distance matrix; can be either a matrix or a dist

n

The number of items in the core

preselected

An optional list of preselected accessions to be included in the core collection; the provided accessions should occur in the labels or rownames of the provided distanceMatrix

coreSelectMethod

The method for computing core accessions within the groups: A-NE (accession nearest entry), E-NE (entry nearest entry) or E-E (entry entry)

adjustedGroupMethod

The method to handle adjusting groups when multiple preselected accessions occur within a single group: split to just split the initial groups with multiple accessions or recompute to recompute the division of accessions over the groups.

algorithm

Algorithm applied to compute a solution: currently, only randomDescent is available

seed

The seed used when generating the core collection. If no seed is provided, a random seed is chosen and each time the recompute() method is called on the object, a new seed will be used.

Details

Based on a provided distanceMatrix and required number n of accessions within the core, a random set of accessions is created, implicitly dividing the full population into initial groups based on the nearest randomly chosen random accession. If a set of preselected accessions is provided, this initial division is adjusted using the adjustedGroupMethod. Then, using the coreSelectMethod in the algorithm, the core accessions within these groups are calculated, resulting in the final core collection.

Fields

adjustedBasedGroups

A list describing the initial random division of all accessions into groups, adjusted for the set of preselected accessions by using the defined adjustedGroupMethod.

adjustedGroupMethod

The method to handle adjusting groups when multiple preselected accessions occur within a single group.

adjustedSelected

A data.frame representing the intial random selection of accesions, adjusted for the set of preselected accessions by using the defined adjustedGroupMethod, with the accession names as labels and the following columns:

  • contains: the (positive) number of accessions that have this accessions as the closest random selected accession

  • preselects: the number of these closest accessions that were preselected

  • preselected: a boolean indicating if the random selected accession was preselected

  • random: a boolean indiciating if the selected accesion was initially randomly chosen or introduced later by the applied adjustedGroupMethod.

algorithm

The applied algorithm to compute the solution.

core

A data.frame representing the core collection with the accession names as labels and in the first and only column a boolean value indicating whether or not the accession was preselected.

coreSelectMethod

The applied method to select the core accessions based on the computed adjustedBasedGroups.

distanceMatrix

The distance matrix; this will allways be a dist object.

n

The required core size

pop

A data.frame representing the whole collection with the accession names as labels and in the first and only column:

  • result: a string describing if the accession is marked as other or as included in the core, and if in the core because it was preselected or because of the applied coreSelectMethod.

preselected

The list of preselected accessions.

randomBasedGroups

A list with the initial division into groups based on the initial random selection of accessions described by randomSelected. Each item describes all accessions that have the random selected accesion from the label as the nearest neighbour, including the random selected accession.

randomSelected

A data.frame representing the intial random selection of accesions with the accession names as labels and the following columns:

  • contains: the (positive) number of accessions that have this accessions as the closest random selected accession

  • preselects: the number of these closest accessions that were preselected

  • preselected: a boolean indicating if the random selected accession was preselected

  • random: a boolean indiciating if the random selected accesion was randomly chosen. This will always be TRUE for this field, but including this column makes the output comparable with adjustedSelected.

seed

The last applied seed for the randomizer. This will only change when the recompute() method is called and no initial seed is defined.

Methods

alternativeCore(n)

The nth alternative core with n a positive integer. Provides for each accession in the core, if available, the nth nearest accession from within the same group as an alternative.

clone(deep = FALSE)

The default R6Class clone method.

initialize(distanceMatrix, n, preselected, coreSelectMethod, adjustedGroupMethod, algorithm, seed)

Initialisation of the object, is called automatically on creation or recomputing.

measure(coreSelectMethod)

The measure for the provided coreSelectMethod. If no value is provided, the current selected coreSelectMethod is used. The measure is used by the algorithm to compute the core collection.

measures()

A data.frame with the available coreSelectMethods as labels and in the first and only column the measures for these methods.

recompute()

Recompute the core collection: If on initialisation of the object a seed was provided, this same seed will be applied and therefore the same core collection will be created. Otherwise, a new seed is generated, resulting in a new core.

print()

Create a summary of the core collection object, same as summary().

summary()

Create a summary of the core collection object, same as print().

See Also

Other core collection: coreCollection-package

Compute selection - recompute method

Description

The function computeAdjustedSelectionUsingRecomputeMethod is used internally by the CoreCollection object to compute an adjusted selection using the recompute method.

Usage

computeAdjustedSelectionUsingRecomputeMethod(dist, adjustedSelected)

Arguments

dist

distance matrix, used for distances and implicitly defining the set of entries

adjustedSelected

the selected entries defined as a list of zero-based integers referring to the row/columns of dist

Details

This function returns a list describing for each of the row/columns entries of dist the closest selected entry. The entries are implicetly defined by the row/columns of dist and referred to by a zero-based integer describing the position.

Compute selection - split method

Description

The function computeAdjustedSelectionUsingSplitMethod is used internally by the CoreCollection object to compute an adjusted selection using the split method.

Usage

computeAdjustedSelectionUsingSplitMethod(dist, groups, preselected)

Arguments

dist

distance matrix, used for distances and implicitly defining the set of entries

groups

the initial division into group defined as a list of zero-based integers referring to the row/columns of dist

preselected

the set of preselected entries

Details

This function returns a list describing for each of the row/columns entries of dist the corresponding entry referred to in groups. However, groups with one or multiple preselected entries are divided, and the returned list wil contain references to the closest preselected entry within this group, implying a split if multiple preselected entries occur within one group. The entries are implicetly defined by the row/columns of dist and referred to by a zero-based integer describing the position.

Compute the core

Description

The function computeCore is used internally by the CoreCollection object to compute the core.

Usage

computeCore(algorithm, method, dist, groups)

Arguments

algorithm

applied algorithm to find solution with method: currently, only randomDescent is available

method

required method for choosing the entries within the groups: A-NE (accession nearest entry), E-NE (entry nearest entry) or E-E (entry entry)

dist

distance matrix, used for distances and implicitly defining the set of entries

groups

the initially created subdivision into groups

Details

The A-NE method requires the core to minimize the average distance between each accession and the nearest entry within the core. The E-NE method requires the core to maximize the average distance between each core entry and its nearest neighbouring entry within the core. The E-E method requires the core to maximize the average distance between all core entries.

Compute a random selection

Description

The function computeRandomSelection is used internally by the CoreCollection object.

Usage

computeRandomSelection(dist, requiredN, preselected, seed)

Arguments

dist

distance matrix, used for distances and implicitly defining the set of entries

requiredN

the required size of the random selection

preselected

a list of preselected entries, referring to the row/column of dist

seed

the applied seed for the randomizer

Details

This function returns a random selection of approximately size requiredN by choosing entries sequentually and randomly, while excluding all entries within a certain radius of an entry chosen before, and by finding iteratively the most appropiate radius to end up with a number close to requiredN of selected entries.